ByTheWay: Boost Your Text-to-Video Generation Model to Higher Quality in a Training-free Way

TL;DR

ByTheWay is a training-free method that enhances text-to-video generation quality by reducing temporal attention disparities and amplifying motion energy, leading to more consistent and dynamic videos without extra training or parameters.

Contribution

It introduces a novel training-free approach that improves structural plausibility, temporal consistency, and motion richness in text-to-video models by leveraging attention map analysis.

Findings

Improves temporal consistency and structural plausibility.

Enhances motion amplitude and richness.

Operates without additional training or parameters.

Abstract

The text-to-video (T2V) generation models, offering convenient visual creation, have recently garnered increasing attention. Despite their substantial potential, the generated videos may present artifacts, including structural implausibility, temporal inconsistency, and a lack of motion, often resulting in near-static video. In this work, we have identified a correlation between the disparity of temporal attention maps across different blocks and the occurrence of temporal inconsistencies. Additionally, we have observed that the energy contained within the temporal attention maps is directly related to the magnitude of motion amplitude in the generated videos. Based on these observations, we present ByTheWay, a training-free method to improve the quality of text-to-video generation without introducing additional parameters, augmenting memory or sampling time. Specifically, ByTheWay is composed of two principal components: 1) Temporal Self-Guidance improves the structural plausibility and temporal consistency of generated videos by reducing the disparity between the temporal attention maps across various decoder blocks. 2) Fourier-based Motion Enhancement enhances the magnitude and richness of motion by amplifying the energy of the map. Extensive experiments demonstrate that ByTheWay significantly improves the quality of text-to-video generation with negligible additional cost.